Just working through some ideas for how to design and assembly parts for a sheet metal drive. I’ve never really understand sheet metal designs that mimic other frames with two independent side rails separated by standoffs, so I took a shot at something that houses all of the wheels in a single part with a repackaged SuperShifter bolted on.
Very cool. One thing to keep in mind is the effect of cascading bend tolerances, but this is only two bends so I can’t imagine it will be that bad.
Yes – that’s been the thing I’ve been most worried about when looking at this sort of part. In this case, I’ve made the center wheel directly driven by the gearbox output and it uses a regular-length hex output, so I think I can get away without supporting it with a bearing pressed into the sheet metal part. So the only things that need to line up will be the two dead axles on each end – I think they’d be able to tolerate misalignment a little bit nicer than a shaft and bearing would.
I would be careful on how you do your lower flanges on both plates.
The lack of a lower flange on the outer plate makes you vulnerable to bending along that whole edge. It could easily cause havoc to any bearings you may have placed in the plate.
Additionally, cutting your flange on the inner plate for the gearbox has a similar effect. You lose a lot of rigidity in the whole drive structure because that length of plate now has a nice spot to bend. Imagine taking a load straight through your direct driven axle and how that area will deflect with or without the flange.
-Parker
The frame as a whole takes some steps to mitigate those potential issues. I didn’t upload it today because the only shot I have of it available at work doesn’t provide a lot of good information. The only drive components that’ll pass through the outer plate are the two dead-axle bolts. There aren’t any bearings or anything that requires very precise alignment.
The belly pan will connect to the outer face in a few places to counteract bending. It’ll also pick up the inner flange and the gearbox flange. That should help to tie it all together.
If I can render a more useful view tonight, I’ll put that up here. The model isn’t anywhere near a finished state yet, but taken as a whole, I think it’ll be pretty stiff.
This looks pretty cool.
I’ve designed enough complete sheet metal frames that I’m comfortable critiquing this.
The first thing I usually look for in a frame is torsional stiffness. It’s hard to tell without seeing the rest of the frame but I think you’ll probably have some torsional stiffness issues if you don’t add a flange to the outside bottom of the wheel bay going either in or out. How you cap the ends of that drivepod is paramount. You’ll need flanges from the outside crossmembers picking up the outside, top and bottom on each end. The giant flange cut out in the middle for the gearbox is also very detrimental unless you can bridge that gap with the gearbox itself. In general putting even a small bend (think 1/4") in a sheet metal part will make it many times stronger. Even after all that I would still put 2-4 box pieces within the wheel bay between the wheels riveted in through the top and sides. That would stiffen this right up.
Some of this relies on your team resources.
-Are you having your parts bent for you or are you going to have to bend them yourself? It would be pretty tricky to bend this part in a bend break, my team hand forms all of our parts and would probably not be able bend this part with the precision needed. You also probably don’t want to try to control the width of your gearboxes and hole alignment with a flange if it’s not bent by a machine.
-How thick of metal is this going to be made out of? If you do the above you should be able to use .060” or .050”.
-Another consideration would be chamfering the corners by the wheels. Otherwise as the metal thicknesses stack up there it becomes easier to get stuck on field elements. On a field with terrain this also allows the wheel to contact the slope you’re trying to go up rather than the frame.
-It looks like you haven’t decided how to attach the outside axles. If you have the resources to tap .5” aluminum rod I would suggest doing that and bolting into it from both sides. This then acts as another piece of structure and helps pull the frame into alignment rather than being a liability like a like axle there probably would.
-You can change all of your 1” flanges to .75” and save a bit of unneeded material and weight.
-Chain tensioning is also a big question…
-How much is your wheel dropped and are you making it adjustable?
Two independent side rails is a good idea that many teams use (including us.) Holding the outside onto the inside with only standoffs sounds like a really bad idea though. I made a drivetrain that is kind of similar to what you’ve got going on about a year ago. If you have any interest I can send you the CAD model.
Neat, sort of reminds me of 971’s frame from this past season. I’m guessing that’s probably similar to what the rest of the frame looks like?
I’m attaching the terrible picture that doesn’t show anything just in case in might show something. 
I appreciate the advice since I’ve never done sheet metal design in robotics applications – but I do have a fair bit of experience with it otherwise.
The crossmembers on each end currently pick up the inner vertical flanges, the outer face and the top surface.
Everything will be laser cut and bent on a CNC break.
-How thick of metal is this going to be made out of? If you do the above you should be able to use .060” or .050”.
Planning on .060" for as much as possible.
-Another consideration would be chamfering the corners by the wheels. Otherwise as the metal thicknesses stack up there it becomes easier to get stuck on field elements. On a field with terrain this also allows the wheel to contact the slope you’re trying to go up rather than the frame.
Chamfering the leading and trailing edges would make the belly pan part a bit trickier, but is completely achievable. Depends on what sort of field we get.
-It looks like you haven’t decided how to attach the outside axles. If you have the resources to tap .5” aluminum rod I would suggest doing that and bolting into it from both sides. This then acts as another piece of structure and helps pull the frame into alignment rather than being a liability like a like axle there probably would.
We’ve done this in the past but usually favor using bolts for axles. We might be able to swing making parts on the lathe since we won’t have to worry about as much manufacturing as we have in the past.
-You can change all of your 1” flanges to .75” and save a bit of unneeded material and weight.
True story.
-How much is your wheel dropped and are you making it adjustable?
That has 3/16" drop right now; it’s dumb and will get changed to 1/8". It isn’t adjustable. I’ll attach another picture that shows a rough idea for chain tensioners.
Two independent side rails is a good idea that many teams use (including us.) Holding the outside onto the inside with only standoffs sounds like a really bad idea though. I made a drivetrain that is kind of similar to what you’ve got going on about a year ago. If you have any interest I can send you the CAD model.
One of my students had a sheet metal drive done over the summer – http://www.chiefdelphi.com/media/photos/38121 – and it’s more typical in its design. I’d like to reduce the part count if possible to make life a bit easier for everyone.
http://imageshack.us/a/img805/4923/70484710152310683625082.jpg
http://imageshack.us/a/img195/8903/70471710152310655105082.jpg
http://imageshack.us/a/img834/4536/70421510152310689040082.jpg
Beautiful CAD. I love seeing sheet-metal designs, if only because our team doesn’t have the resources to ever do one on the level that we’d like to (and I don’t think we have the patience to CAD everything :P).
I do have a question regarding the custom housing for the gear-box, namely, why? Don’t get me wrong, I’m not criticizing the design-- I don’t have nearly enough experience with custom gearboxes to offer anything constructive or unique there-- but I assume it’s for weight? If so, then what are the actual weight savings over the original housing?
Hi Madison,
Your tensioner brackets look a little flimsy. Try a triangular shape instead of the long rectangular flanges. The triangle will distribute the force to the base of the flange instead of bending or twisting. The forces on a collision and the stress of changing directions is directly on the chain tensioner. It needs to be strong and also think of putting two slots on the top flange instead of thur holes and insert the tensioner from the top. Use 3 screws to locate and tighen down. That way you could shim the tensioner and that will adjust the tension on the chain by moving the tensioner a little higher. I think the tensioner should go in the middle between the wheels. When the chain starts to stretch it will sag on the top and might possible rub together where you have it.
One important thing is you need to take a 1:1 cutaway of the main piece and make sure the goose neck brake die will fit inside the part and clears the first flange when the last bend takes place. If the brake die isn’t the right shape you can’t make the last bend without hitting the other flanges. This results in what they call a back bend. You could plot the flat pattern on paper and fold the plot up and experiment with where you want the flanges. Paper and tape is a sheet metal designers tools.
I wouldn’t worry too much about the cutout on the bottom for the gear box. Once you have it riveted to the belly pan it will lock down and stiffen.
There are some places were you can tab and notch the sheet metal for locating and increasing the strength. Ask your sheet metal sponsor how you can do it. Basically it is like a creating a jig saw piece where the two pieces fit and lock together.
On the gear box plates you might consider to bend the side plates flanges the opposite direction to clear the chain and sprocket.
Your front and back panels would be stronger if you added some triangular gussets when you assemble the frame. Also could add another return flange to the top flange.
Need chain brake access. Your design looks pretty tight inside where the wheels are. Why?
Think about how you want to mount your battery and build a cool bracket that rivets on and holds your reset switch.
There are pictures of our 2010 soccer playing robot. Your frame shape is very similar to that one. It was a 8 wheeler with idler sprockets to allow for fast traverse on the big bump with the lowest COG
There are two reasons for repackaging a gearbox. The first reason is for form factor. Repackaging a gearbox means that you can make your frame adapt to it in any way you want. The second reason is weight. The repackaged gearbox has walls that are thin(comparably to the original housing). The Supershifter walls are .25" thick, the repackaged ones are at most .125" thick. Assuming that the sidewall sizes are the same, you basically just removed the weight of one AM shifter sidewall.
Andymark Shifter Sidewall = 5" * 5" * .25" * .0975 lbs/in^3(density of 6061 al) = .61 lbs
So you saved .61 of a pound if not more.
Thank you very much for that response, Michael-- very helpful.
Team 2220 might have to consider doing custom gearboxes this next year (though we’ll probably still be a 120lb robot anyway).
All of this is true, though the sheet metal shown is .060" rather than .120"
The real reason I did this, though, was because on a stock SuperShifter, the output shaft is 1.5" from the closest edge and, with 4" wheels, that leaves very little clearance between the Super Shifter (and its servo) and the ground. By repackaging it, I was able to increase the ground clearance – though it seems to require pneumatic shifting rather than servo shifting.